help-glpk
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Help-glpk] Intermediate solutions


From: Carlos Javier Sosa Gonzalez
Subject: [Help-glpk] Intermediate solutions
Date: Mon, 23 Jun 2003 23:16:35 +0100 (WET DST)

Hello,

I have an iterative algorithm that formulates an MIP problem in each
algorithm iteration. Each MIP problem has an optimum solution, that it is
reached after several seconds. I have looked at the glpsol output and the
glpsol gives some non-optimal solutions before finishing with the optimal.
For example:

Executing file 5xp1.pla.7.cplex!
lpt_read_prob: reading LP data from `5xp1.pla.7.cplex'...
lpt_read_prob: 234 variables, 927 constraints
lpt_read_prob: 1197 lines were read
gm_scal: max / min = 1.000e+00
gm_scal: max / min = 1.000e+00
lpx_adv_basis: size of triangular part = 927
      0:   objval =   0.000000000e+00   infeas =   1.000000000e+00 (0)
    197:   objval =   7.000000000e+00   infeas =   1.099025249e-18 (0)
*   198:   objval =   7.000000000e+00   infeas =   0.000000000e+00 (0)
*   200:   objval =   7.000000000e+00   infeas =   0.000000000e+00 (0)
*   390:   objval =   2.000000000e+01   infeas =   2.161255839e-15 (0)
OPTIMAL SOLUTION FOUND
Integer optimization begins...
+   390: mip =     not found yet; lp =   2.000000000e+01 (1, 0)
+   628: mip =   1.600000000e+01; lp =   2.000000000e+01 (8, 7)
+   835: mip =   1.700000000e+01; lp =   2.000000000e+01 (13, 14)
+  2006: mip =   1.700000000e+01; lp =   2.000000000e+01 (19, 56)
+  3813: mip =   1.700000000e+01; lp =   2.000000000e+01 (40, 101)
+  5477: mip =   1.700000000e+01; lp =   2.000000000e+01 (9, 180)
+  5512: mip =   1.700000000e+01; lp =     tree is empty (0, 189)
INTEGER OPTIMAL SOLUTION FOUND
CPU:              4880000       670000

The questions are:

* Are these non-optimal solutions of the problem? or can they violate any
problem constrain or bound? are they feasible solutions?

* If they are solutions of the problem, how can I access to these
MIP non-optimal solutions?

The reason is that these points will be optimum points in future
iterations of my algorithm.


Thanks in advance,
Javier Sosa

--
 Carlos Javier Sosa González  E.T.S.I. de Telecomunicación
  address@hidden         Campus Universitario de Tafira, Pab. A, Desp. 301 
  Tel.: +34 928 457 324       Univ. de Las Palmas de Gran Canaria
  Fax : +34 928 451 243       35017 Las Palmas de Gran Canaria
  ICQ : 1756350985            Canary Islands. SPAIN






reply via email to

[Prev in Thread] Current Thread [Next in Thread]